The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute logic from the @nomadamas/k-skill npm package via npx and runs scripts/install.sh, which performs shell operations and repository cloning.
[EXTERNAL_DOWNLOADS]: The scripts/install.sh script clones a remote repository from github.com/kimlawtech/korean-privacy-terms during the installation phase.
[PERSISTENCE]: The skill modifies ~/.claude/skills/ and ~/.agents/skills/ directories, which are standard paths for agent skill discovery and execution, to install and update "upstream" components.
[DYNAMIC_EXECUTION]: The skill generates React/Next.js components and MDX files at runtime based on templates and user-supplied information.
[INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: User input collected during the "Mandatory interview" phase (e.g., service details, business information) as defined in instruction.md. 2. Boundary markers: The instructions require the inclusion of legal disclaimer comments in the generated output files. 3. Capability inventory: The skill utilizes git, rsync, cp, and npx to fetch and deploy files, and performs file-write operations within the project directory. 4. Sanitization: No explicit sanitization or escaping of user-provided information is mentioned before it is interpolated into the generated code files.

korean-privacy-terms